Hydrogen Bonding and Jahn–Teller Distortion in Groutite,α-MnOOH, and Manganite,γ-MnOOH, and Their Relations to the Manganese Dioxides Ramsdellite and Pyrolusite

JOURNAL OF SOLID STATE CHEMISTRY 133, 486—500 (1997)ARTICLE NO. SC977516

Hydrogen Bonding and Jahn–Teller Distortion in Groutite, a-MnOOH,and Manganite, c-MnOOH, and Their Relations to the Manganese

Dioxides Ramsdellite and Pyrolusite

Thomas Kohler and Thomas Armbruster1Laboratorium fu( r chemische und mineralogische Kristallographie, Universita( t Bern, Freiestrasse 3, CH-3012 Bern,Switzerland

and

Eugen LibowitzkyDivision of Geological and Planetary Sciences, California Institute of Technology, Pasadena, California 91125

Received December 16, 1996; in revised form June 18, 1997; accepted June 23, 1997

The crystal structures of groutite, a-MnOOH (space groupPnma, a510.667(1), b52.871(1), c54.554(1) A_ , Z54), andmanganite, c-MnOOH (space group P21 /c, a55.304(1),b55.277(1), c55.304(1) A_ , b5114.38(2)°, Z54), both from theKalahari manganese field (South Africa), were refined includinghydrogen positions from room temperature X-ray single-crystaldata. The refinements converged to R values of 1.5% for 479(groutite) and of 2.0% for 821 (manganite) unique reflections,respectively. A (101) twin refinement based on F2

obs was appliedfor manganite leading to a twin contribution of ca. 0.9 : 0.1. Thestructures of groutite and manganite are distorted derivatives ofthe MnO2 polymorphs ramsdellite and pyrolusite (rutile struc-ture), respectively. The structural distortions of the oxyhydrox-ides are caused by an interaction of Jahn–Teller distortion ofoctahedrally coordinated Mn31 (four short and two long Mn–Odistances) and hydrogen bonding. In both structures two symmet-rically distinct O sites (O1 and O2) are three-coordinated byMn31. The choice which O site forms an OH group is governedby the orientation of the Jahn–Teller distortion and space con-straints dictated by the octahedral framework topology. The twolong Mn–O bonds are formed by both O1 and O2 thus theJahn–Teller distortion alone does not determine the preferenceof the OH group. In groutite, the OMn3 coordination fragmentwhich shows the strongest deviation from planarity toward a tri-gonal pyramid bonds to H where the O–H vector points perpen-dicular to the Mn3 plane. In manganite the coordinations of O1and O2 are very similar, thus H shows long range disorder asobserved by twinning. Both, groutite and manganite have shortO–H···O distances (2.6 A_ ) giving rise to peculiar IR absorptionfeatures between 3200 and 1800 cm21. Oriented single-crystalslabs of manganite and diaspore (isostructural to groutite) werestudied by polarized FTIR spectroscopy at 82 and 298 K, and all

1To whom correspondence should be addressed.

480022-4596/97 $25.00Copyright ( 1997 by Academic PressAll rights of reproduction in any form reserved.

H-bonding related IR-absorptions were assigned. Manganite,c-MnOOH, and groutite, a-MnOOH, both transform in airabove ca. 300°C to b-MnO

2(pyrolusite) which was studied by

in situ temperature dependent single-crystal X-ray diffraction.The topotactic relation is preserved during the transformation.( 1997 Academic Press

INTRODUCTION

The group of oxyhydroxides M3ÒOH may be dividedinto two structural families; one with octahedral coordina-tion and strong hydrogen bonds and another where M isseven-fold coordinated without strong hydrogen bonds. Inthe latter family of structures (M3`"Ho, Tb, Yb, Lu, Er,Y) six oxygen atoms, three of which are hydroxyl groups,form a trigonal coordination prism around M, and theseventh O resides along a normal to a prism face closest tothe M center. Oxyhydroxides with octahedrally coordinatedM are found for M3`"Al, Sc, Y, V, Cr, Mn, Fe, Co, Ni, Rh,Ga, and In. It is one of the characteristics of these com-pounds that various modifications exist, commonly labeledby a Greek letter as prefix. Unfortunately, the same prefix isused for different structure types depending on the chemicalcomposition. Two modifications with octahedrally co-ordinated Mn3 , groutite, a-MnOOH, and manganite,c-MnOOH, are subject of this study. Both structures arecomposed of edge-sharing Mn3Ò

6octahedra which are

further linked to a three-dimensional framework by sharingvertices (Fig. 1). Edge-sharing single chains are found forderivatives of the rutile structure (e.g., manganite), whereasedge-sharing octahedral double chains are characteristic ofa-AlOOH, diaspore; a-FeOOH, goethite; a-MnOOH, grou-tite. The same arrangement of octahedral double chains is

6

FIG. 1. Comparison of the arrangement of coordination octahedra inMnO

2and MnOOH polymorphs. H atoms in MnOOH polymorphs are

shown by small spheres with hydrogen bonds.

2Two crystals ‘‘have the same arrangement’’ and are isotypic (or iso-structural) if their crystal structures are congruent with respect to spacegroup and occupancy of equivalent positions (8). In principle, these struc-tures (c-MnOOH, InOOH, and b-CrOOH) are very similar concerning thearrangement of M3` and O but not with respect to the hydrogen positions.In decision to this fact, these structures are called isotopological, a termoften used in zeolite crystal chemistry.

HYDROGEN BONDING IN GROUTITE AND MANGANITE 487

also known for MnO2, ramsdellite (Fig. 1). Manganite,

c-MnOOH, possesses also a structurally related MnO2

modification named pyrolusite, b-MnO2

(rutile structuretype; Fig. 1). Feitknechtite, b-MnOOH, a third modifica-tion, was not considered in this study because the structureis not well defined and suitable single crystals are notavailable.

Manganese dioxides and oxyhydroxides are of consider-able importance in many technical applications. In alkalinebatteries, for example, c-MnO

2, related to the mineral nsu-

tite, forms the basis due to specific physical and chemicalproperties. A detailed review on this subject is provided byChabre and Pannetier (1). The structure of c-MnO

2has

been described first as a random intergrowth of pyrolusiteand ramsdellite layers (2) and was then shown to incorpor-ate microtwinning defects (1). It has been proposed (3) thatthe second kind of defect does not occur in reducedMnOOH compounds, which means that reduction in batte-ries leads to a ‘‘de-microtwinning’’ of c-MnO

2and to the

formation of an intergrowth of manganite and groutite.During the electrochemical processes, Mn4` is partly re-duced to Mn3`, thus MnO

2transforms to Mn

2O

3or

MnOOH. The mechanism of reduction may be differentbecause the modifications show a wide range of phase

boundaries and are able to absorb protons with only minorstructural changes. New developments show that highlycrystalline a-MnO

2phases (hollandite-type) have high

performances as electrode materials for secondary lithiumcells (4).

The octahedral framework of manganite and pyrolusite ischaracterized by endless 1]1 channels (Fig. 1) where inmanganite the hydrogen atoms reside. A rough structuremodel of manganite was proposed by Garrido (5) whoconsidered manganite a marcasite-like structure. A sub-sequent structural reinvestigation showed that manganite isactually monoclinic and only pseudo-orthorhombic (6). Theapparent orthorhombic symmetry is caused by polysyn-thetic twinning parallel to (101) in P2

1/c setting. Dachs (7)

derived the H position from a neutron diffraction study andshowed the hydrogen bonding scheme. Isotopological2 tomanganite are InOOH and b-CrOOH. The hydrogen posi-tion in orthorhombic InOOH (space group P2

1nm, a"

5.26, b"4.56, c"3.27 A_ ) was determined by neutron dif-fraction (9), yielding a H arrangement different to manga-nite. b-CrOOH is isotypic to InOOH and has a"4.588,b"4.292, c"2.955 A_ (10).

In ramsdellite (11) and groutite (12) the connectivity ofthe double chains leads to 2]1 channels in the framework(Fig. 1) where in groutite the hydrogen atoms are located.Gruner (12) noted that groutite (a-MnOOH) is isostructuralto goethite (a-FeOOH) and diaspore (a-AlOOH). A struc-ture refinement neglecting H positions was performed at-tributing the large variations in Mn—O bond lengths to theJahn—Teller effect (13). Using neutron diffraction Busingand Levy (14) located the H atom in diaspore, a-AlOOH,and noticed that the H atom is not on the connecting line ofO atoms. Instead the OH bond makes an angle of 12.1(3)°with the O—O vector.

Since the most recent studies on groutite (13) and manga-nite (7), X-ray data acquisition and correction procedureshave become significantly improved. In addition, structuresof twinned crystals can be refined from single-crystal data.Thus, it was an aim of this paper to locate hydrogen atomseven in these transition metal oxyhydroxides with X-raymethods. The relatively short O—H···O distances (ca. 2.6 A_ )in MnOOH polymorphs give rise to characteristic IR ab-sorptions which are not yet well understood. Thus, orientedsingle-crystal IR absorption spectra were collected andinterpreted. Furthermore, temperature dependent cell di-mensions as well as topotactic relations between MnOOH

488 KOHLER, ARMBRUSTER, AND LIBOWITZKY

and MnO2

polymorphs were investigated by in situ high-temperature X-ray diffraction studies of groutite and man-ganite under oxidizing conditions.

EXPERIMENTAL

X-ray Data Collection and Structure Refinement

Manganite and groutite from the Nchwaning Mine in theKalahari manganese field (Republic of South Africa) haveend-member compositions, as verified by electron micro-probe analyses. Selected crystals of manganite (150]100]250 lm) and groutite (100]70]350 lm) from this localitywere mounted on glass fibers and transferred to an EnrafNonius CAD4 diffractometer (graphite-monochromatedMoKa radiation) for structural study at room temperature.Experimental details for data collection and structure re-finements are given in Table 1. Intensity data were empiric-ally corrected for absorption (t scans). Data reduction,including background and Lorentz polarization correc-tions, was carried out using the SDP program library (15).Neutral atom scattering factors were used for the structurerefinements.

Starting parameters for manganite refinements were ad-opted from Dachs (7) but the B-lattice setting was trans-formed to the primitive standard setting P2

1/c using the

transformation matrix 120 12; 0!1 0; !1

20 1

2.

TABLE 1Crystal and Refinement Parameters for Groutite and Manganite

Groutite Manganite

Crystal size 0.1]0.07]0.35 mm 0.15]0.1]0.25 mmSpace group Pnma P2

1/c B2

1/d

a (A_ ) 10.667(1) 5.304(1) 8.915(1)b (A_ ) 2.871(1) 5.277(1) 5.277(1)c (A_ ) 4.554(1) 5.304(1) 5.748(1)b (deg) 114.38(2) 90.02(1)Z 4 4 8Volume (A_ 3) 139.47 135.2 270.4Twinning contribution 0.9:0.1

Data Collection and Refinement ParametersScan type u (deg)#0.4 tan h 1.5 1hmax (deg) 40 42No. of measured reflections 3468 3378No. of unique reflections 494 876No. of observed reflections [F

0'4p(F

0)] 479 821

No. of parameters 24 34R (%) 1.47 2.04R

8(%) 1.60

wR2

(%) 4.85Goof"S 1.7904 1.096

Note. R"(+ D[F(0"4)

!F(#!-#)

] D )/(+F(0"4)

);

R8"M(+(weight D[F

(0"4)!F

(#!-#)] D )2)/(+(weightF2

(0"4)))N1@2;

wR2"M[+weight(F2

(0"4)!F2

(#!-#))2]2/[+weight(F2

(0"4))2]N1@2;

Goof"M(+(weight D[F(0"4)

!F(#!-#)

] D )2)/(reflections0"4

!parameters(3%&)

)N1@2.

A preliminary manganite refinement in space group P21/c

led to an R value of 4.8%. However, this result had to bedisregarded because many symmetry forbidden reflectionswith highly significant intensities were observed. These re-flections are due to (101) twinning as already noticed byDachs (7). A second structure refinement based on F2

0"4with

the program SHELXL93 (16) was performed for a twinnedmanganite model, resulting in an R value of 2.0%. Thehydrogen position was determined from difference Fouriermaps and subsequently refined. The largest residual peaksin these maps were $1.35 e/A_ 3 surrounding the Mn atoms.

Groutite starting parameters were taken from DentGlasser et al. (13) and refined with the program SHELXTL(17) using a 1/p2 weighting scheme, converging at R"

1.5%. The hydrogen position was determined from differ-ence Fourier maps and subsequently refined. The largestresidual peaks in these maps were $0.53 e/A_ 3.

X-Ray Investigation of Manganite and Groutiteat Elevated Temperatures

The same manganite crystal as used for the single-crystaldata collection at room temperature was remounted ona SiO

2glass fiber with a high temperature resistant resin.

The specimen was transferred to the Enraf Nonius diffrac-tometer equipped with a high temperature device where thecrystal was heated by a temperature-controlled hot airstream. Starting at 25°C, the temperature was successivelyincreased in steps of ca. 25 to 325°C. At each step, thetemperature was kept constant for 1 h and the cell dimen-sions were refined. The crystal was kept for 18 h at 175°C totest the thermal stability.

This procedure was repeated with an additional groutitecrystal (200]170]350 lm) which was heated up to 350°C.However, the thermal stability test at 175°C was not per-formed.

FTIR Spectroscopy

FTIR spectra were recorded on a Nicolet 60SX FTIRspectrometer using a glowbar light source, a KBr beamsplitter, and a liquid nitrogen cooled MCT detector. For thelow-temperature spectra (82 K) a commercial cryocell(MMR Technologies) with KBr windows was used. Polariz-ed single-crystal spectra of manganite and diaspore weremeasured using a gold wire grid polarizer on AgBr substrate(extinction ratio at 2000 cm~1&1 : 100) and a sample aper-ture of 400 lm. For the measurements of an extremely small((200 lm diameter) manganite (010) and diaspore (001)slab, and for testing of different, small manganite twindomains, a Nicplan FTIR microscope with a 15]/0.58 N.A.cassegrain objective (Spectra-Tech Inc.) was employed.Spectra were recorded at a resolution of 2 cm~1 (4 cm~1


with the microscope) and were averaged from 256 scans forthe powder data and from at least 1024 scans for thesingle-crystal data.

Manganite and groutite powder samples were preparedby the conventional KBr pellet technique. However, sincespectra from KBr powder pellets frequently show artifactslike impurities from trace amounts of water, surface OHgroups, or organic substances, single-crystal samples werealso prepared. Manganite and groutite crystals wereoriented according to the morphology and cleavage of thecrystals, whereas for diaspore, X-ray oriented, gem-qualitycrystal slabs from Shannon et al. (18) were available. Theoriented and cut slabs were ground and polished on bothsides until a final thickness of 2—6 lm was obtained. A de-tailed description of the polishing procedure is given byLibowitzky and Rossman (19). Due to the small size (only&300 lm) and perfect cleavage, thin groutite single-crystalslabs could not successfully be prepared, and thus the iso-typic Al-oxyhydroxide was used.

RESULTS

X-Ray Structure Refinement of Groutite

The cell dimensions of groutite in space group Pnma area"10.667(1), b"2.871(1), c"4.554(1) A_ . Refined atomiccoordinates and atomic displacement parameters are givenin Table 2. Selected interatomic distances and angles forgroutite are given in Table 3. The oxygen atoms formdistorted octahedra around the manganese atoms. Fourshort Mn—O bonds form a rough square (2]O1 1.895(1)and 2]O2 1.965(1) A_ ) while two long bonds, Mn—O1(2.174(1) A_ ) and Mn—O2 (2.338(1) A_ ), complete the octahed-ron. The octahedral edges are shared to form double chainsparallel to b. In the resulting tunnels, the hydrogen atom isbonded to O2 within 0.81(4) A_ to form an OH group witha hydrogen bond to O1 within 1.82(4) A_ .

TABLRefined Atomic Coordinates and Atomic Displac

Atom x/a y/b z/c º11

º22

Groutite, spacMn 0.14014(1) 1

4!0.05014(3) 0.00696(9) 0.00534(

O1 !0.18652(7) 14

0.2989(2) 0.0082(3) 0.0067(3O2 !0.07017(7) 1

4!0.1947(2) 0.0087(3) 0.0076(3

H2 !0.099(3) 14

0.641(8)

Manganite, spaMn !0.23684(3) 0.01033(3) 0.75464(2) 0.00451(8) 0.00564(O1 0.3749(1) 0.1238(1) 0.6279(2) 0.0061(2) 0.0062(2O2 0.8752(1) 0.1256(1) 0.1206(2) 0.0058(2) 0.0067(2H1 0.284(4) 0.027(5) 0.725(4)

X-Ray Structure Refinement of Manganite

Single-crystal X-ray data of manganite yielded the celldimensions a"5.304(1), b"5.277(2), c"5.304(1) A_ , b"114.38(2)°, and a (101) twin ratio of 0.9 : 0.1. Refined atomiccoordinates and atomic displacement parameters are givenin Table 2. Selected interatomic distances and angles formanganite are given in Table 3.

The structure of manganite consists of Mn octahedrasharing edges to form single octahedral chains. These chainsare corner-linked (by O1 and O2) parallel to the b-axis toa three-dimensional framework. Mn3` is bonded in a roughsquare twice to O1 with 1.977(1) and 1.982(1) A_ and twice toO2 with 1.881(1) and 1.893(1) A_ , respectively. Two addi-tional long bonds, Mn—O2 (2.213(1) A_ ) and Mn—O1(2.337(1) A_ ), complete the octahedron. The hydrogen posi-tion is located in the interstitial channels between the chainsand bonds to O1 within 0.98(2) A_ , thus forming an OHgroup. In addition, H forms a 1.61(2) A_ hydrogen bondto O2.

Behavior of Manganite and Groutite at ElevatedTemperatures

To elucidate the topotactic relation between manganiteand pyrolusite the B-centered manganite cell setting (7),i.e. B2

1/d with a"8.88, b"5.25, c"5.71 A_ , b"90.0° is

chosen because the cell orientation in this setting is the sameas in pyrolusite (see Discussion). The twin contribution inmanganite is not altered by heat treatment. This is con-firmed by measuring the intensities of the B2

1/d forbidden

reflections characteristic of the twin which survive up to300°C. At 325°C, no intensities for the current cell settingwere found. A subsequent search of 20 strong reflectionsyielded a new orientation matrix and a distorted pseudo-tetragonal cell with a"4.409(7), b"4.421(3), and c"2.871(2) A_ at 25°C. This cell is characteristic of distortedpyrolusite, b-MnO

2. In spite of the strong differences in cell

E 2ement Parameters for Groutite and Manganite

º33

º23

º13

º12

º%2

e group Pnma9) 0.00620(9) 0 0.00099(4) 0 0.00616(5)) 0.0073(3) 0 0.0017(2) 0 0.0074(2)) 0.0058(3) 0 !0.0001(2) 0 0.0074(2)

0.038(8)

ce group P21/c

8) 0.00477(9) !0.00056(3) !0.00209(5) 0.00070(3) 0.00492(6)) 0.0066(2) !0.0004(2) !0.0035(2) 0.0001(2) 0.0060(1)) 0.0064(2) !0.0008(2) 0.0030(2) 0.0007(2) 0.0062(1)

0.029(7)

TABLE 3Selected Interatomic Distances (A_ ) and Angles (Deg) for

Groutite and Manganite

Atom1 Atom2 Dist A—O—B Angle

GroutiteMn O1 1.895(1)

O1 1.895(1)O2 1.965(1)O2 1.965(1)O1 2.174(1)O2 2.338(1)

average O 2.039O2 H2 0.81(4) Mn—O2—H2 129(2)

Mn—O2—H2 112(1) 2]Mn—O2—Mn 101.80(3) 2]Mn—O2—Mn 93.86(3)

O1 H2—O2 2.568(1) Mn—O1—Mn 122.46(2) 2]Mn—O1—Mn 98.52(4)O1—H2—O2 171(4)

ManganiteMn O2 1.881(1)

O2 1.893(1)O1 1.977(1)O1 1.982(1)O2 2.213(1)O1 2.337(1)

average O 2.047O1 H1 0.98(2) Mn—O1—H1 107(1)

Mn—O1—H1 109(1)Mn—O1—H1 92(2)Mn—O1—Mn 124.10(3)Mn—O1—Mn 124.80(3)Mn—O1—Mn 97.63(2)

O2 H1—O1 2.595(1) Mn—O2—Mn 126.86(3)Mn—O2—Mn 125.75(3)Mn—O2—Mn 94.65(2)O1—H1—O2 178(2)

TABLE 4Cell Constants of Groutite and Manganite (B-Centered Setting)

As a Function of Temperature

¹ (°C) a b c

Groutite, Pnma25 10.665(2) 2.866(2) 4.553(3)50 10.671(2) 2.866(2) 4.554(3)75 10.677(2) 2.867(2) 4.558(2)

100 10.682(3) 2.867(2) 4.559(3)125 10.687(3) 2.868(2) 4.559(3)150 10.693(2) 2.868(2) 4.560(3)175 10.698(3) 2.869(2) 4.562(3)200 10.699(3) 2.689(2) 4.562(3)225 10.703(4) 2.870(2) 4.564(3)250 10.712(3) 2.870(2) 4.564(3)275 10.715(5) 2.871(3) 4.567(2)300 10.719(7) 2.872(3) 4.566(4)325 10.733(7) 2.873(3) 4.566(4)% elongation 0.6 0.2 0.3

Manganite, B21/d setting, b"90.0°

25 8.195(1) 5.277(1) 5.748(1)48 8.920(1) 5.282(1) 5.752(1)75 8.923(1) 5.284(1) 5.754(1)

100 8.926(1) 5.287(1) 5.755(1)125 8.926(1) 5.289(1) 5.755(1)150 8.930(1) 5.290(1) 5.756(1)175 8.930(1) 5.292(1) 5.756(1)175 (18 h) 8.929(1) 5.292(1) 5.756(1)200 8.931(1) 5.293(2) 5.757(1)225 8.933(1) 5.296(2) 5.757(1)250 8.934(1) 5.299(2) 5.759(1)275 8.937(1) 5.303(2) 5.761(1)300 8.937(2) 5.305(4) 5.759(2)% elongation 0.2 0.5 0.1


dimensions between manganite and pyrolusite the trans-formation is of topotactic nature where the arrangement ofoctahedral chains remains preserved (Fig. 1). The mostpronounced effect during heating of manganite is the in-crease of the b-axis from 5.277(1) A_ at 25°C to 5.305(4) A_ at300°C (expansion of 0.5%), which after oxidation decreasesto 4.421(3) A_ for b-MnO

2(contraction of 16%). The expan-

sion of a and c of 0.2% and 0.1% between 25 and 300°C isless pronounced, and upon transformation a/2 of manganitecontracts 1% whereas the contraction of c/2 is not signifi-cant.

The intensities of groutite characteristic X-ray reflectionsdecreased strongly at 300°C and faded completely at 350°C.At 350°C only 0k0 groutite reflections (equivalent to 00lreflections in pyrolusite) remained observed. A subsequentsearch for reflections at room temperature also yielded anangularly distorted pseudotetragonal unit cell (a"b"

4.418(2), c"2.862(1) A_ ) characteristic of pyrolusite. How-ever, the pyrolusite reflections were extremely smeared outin the omega directions leading to a half-width of ca. 4.5° forhk0 and 2.5° for 00l reflections, respectively. When groutiteis heated, the a-axis increases from 10.665(2) A_ at 25°C to10.733(7) A_ at 325°C (expansion of 0.6%) and at 350°Ccollapses to ca. a/2"4.42 A_ (contraction of 17%), charac-teristic of a and b of pyrolusite. The expansion of c andb (0.3% and 0.2%) of groutite is less evident, and upontransformation these axes contract by 3% and 0.1%, respec-tively.

The oxidation of Mn3` in manganite and groutite toMn4` in pyrolusite is accompanied by dehydration of thestructure. In case of groutite, oxidation changes the arrange-ment of the octahedral framework but the orientation of theformerly 2]1 channels is maintained in the orientation ofthe 1]1 channels of pyrolusite. A list of cell dimensionswith corresponding temperatures is given in Table 4 andshown in Figs. 2 and 3 for groutite and manganite, respec-tively.

FIG. 2. Cell constants of groutite as a function of temperature.


FTIR Measurements on MnOOH Polymorphs

FTIR powder spectra of groutite and manganite at roomand liquid nitrogen temperatures are presented in Figs. 4and 5. A list of peak positions and band assignments is givenin Table 5. Even though spectra of manganite and groutitehave been published in previous papers (20, 21), the highaccuracy of the FTIR spectroscopy (resulting in a revisedlist of peak positions), the acquisition of low temperaturedata, and, finally, the improved peak assignment in combi-

nation with the discussion of H bonds below, justify andmoreover require the exhibition of the present data. Inaddition, the polarized spectra of manganite (Fig. 6) anddiaspore single-crystals (Fig. 7; applicable also to groutite,goethite, etc.) facilitate an accurate band assignment anddiscussion.

For a complete factor group analysis of all IR-allowedmodes of the OH groups (and their polarization behavior)in manganite and groutite, the VECTOR-tables (22) wereused. In manganite, using the B12

1/d1 setting (factor group

FIG. 3. Cell constants of manganite as a function of temperature.


C2h

), the OH stretching modes (l) belong to species Au(par-

allel to b) and Bu(parallel to a). The in-plane bending modes

(d), which are also parallel to the (001) plane, belong tospecies A

uand B

uwith polarizations parallel to b and a. The

out-of-plane bending mode (c) is exactly perpendicular tothe (001) plane and consequently belongs to species B

upar-

allel to c. The spectra of manganite and groutite show thefollowing features: (a) A broad absorption band around2600—2700 cm~1 which is polarized in the plane of the OHvectors. According to Novak (23) this frequency is in goodagreement with an OH stretching mode belonging to a hy-drogen bond with an O—H···O length of &2.60 A_ (which

is observed in manganite and groutite). (b) A single bandin manganite and a double band in groutite around2000 cm~1 shows the same polarization behavior. This in-teresting feature in the spectra will be the subject of theextensive discussion below. (c) The OH bending modes areobserved around 1000—1150 cm~1. According to the cor-relation diagrams (23), this frequency region is in goodagreement with the frequencies of the stretching modes andthe observed lengths of the H bonds. There are three bandsin manganite and, on a first glance, only two in groutite. Thelowest-energy band of this group is assigned to the c mode(perpendicular to the OH vector plane), and the remaining

FIG. 4. FTIR powder spectrum of groutite at 82 and 298 K. Tinyspikes in the spectra which are caused by organic impurities, uncompen-sated water, and CO

2are marked by asterisks.

TABLE 5Peak Positions (cm~1) of the Groutite and Manganite FTIR

Powder Spectra (from Figs. 4 and 5) at 298 and 82 K

Groutite Manganite

298 K 82 K Mode 298 K 82 K Mode

2850 2855 3 c,d OH? &2900 &2905 3c,d OH?2686 2665 l OH 2660 2627 l OH2026 2045 2 d OH 2060 2083 2c,d OH1932 1951 2 c OH — — —— — — 1151 1160 d1 OH

1027 1035 d OH 1116 1124 d2 OH999 1008 c OH 1086 1090 c OH618 621 l.m. &638 &640 l.m.582 587 l.m. 599 604 l.m.544 549 l.m. 509 513 l.m.

&453 &458 l.m. 453 457 l.m.

Note. ESD’s are $2 cm~1, except where indicated by ‘‘&’’ (diffusepeak or shoulder). l, stretching mode; d, in-plane bending mode; c, out-of-plane bending mode; l.m., Mn—O lattice mode; 2, first overtone; 3, secondovertone.


two (manganite) or one (groutite) represent the d modes inthe OH plane. However, single-crystal spectra of diasporeshow that there is also a second d mode in the diasporegroup spectra (parallel to a). The very weak appearance isexplained by the OH vector orientation which is almostparallel to this axis. Hence, the bending mode (which isperpendicular to the vector) has almost no component par-allel to that axis. (d) The lattice modes (Mn—O vibrations)are observed below 700 cm~1.

FIG. 5. FTIR powder spectrum of manganite at 82 and 298 K. Tinyspikes in the spectra which are caused by organic impurities, uncompen-sated water, and CO

2are marked by asterisks.

In groutite (factor group D2h

) the OH vectors are alignedparallel to the (010) plane. Thus, l belongs to B

1u(parallel to

c) and B3u

(parallel to a). The in-plane bending modesbelong to the same species, whereas for c (which is perpen-dicular to that plane) only the B

2umode (parallel to b) is

active.

FIG. 6. Polarized FTIR spectra of manganite single-crystals. The as-terisk marks a tiny spike in the ‘‘a’’ spectrum which belongs to the bendingmode in b (caused by the conical light path connected with the high N.A. inthe cassegrains of the FTIR microscope).

FIG. 7. Polarized FTIR spectra of diaspore single crystals. Samplethickness &7 lm. Most of the high-intensity peaks are truncated.


DISCUSSION

Structure of Groutite

The structure of groutite, a-MnOOH, is a distorted deriv-ative of ramsdellite, MnO

2, and is isostructural to diaspore,

FIG. 8. The Mn3` oxygen coordination in groutite. Atom displacemenattached to O2 at the apex of the octahedron.

a-AlOOH (14), and goethite, a-FeOOH (24). The elongatedMn3Ò

6octahedron (four short and two long Mn—O

bonds) can be attributed to Jahn—Teller distortion andhydrogen bonding. To determine whether O1 or O2 istopologically preferred for an OH group, the OMn

3frag-

ments in groutite are studied (Fig. 8). The O2Mn3

fragmentshows the stronger deviation from planarity toward a trig-onal pyramid than the O1Mn

3fragment. The average

Mn—O2—Mn angle is 99.15° (2]101.8° and 93.86°) in com-parison to the average Mn—O1—Mn angle of 114.48°(2]122.46° and 98.52°). Therefore, H2 binds to O2, wherethe H2—O2 bond (0.81(4) A_ ) is perpendicular to the Mn

3plane. Decreased O—H distances (0.8 A_ versus ca. 1.0 A_ ) arecharacteristic of hydrogen positions derived from X-raydiffraction data where not the position of the nucleus but ofthe maximum electron density is refined. To fulfill bondvalence requirements for Mn3 , the OH group forms thelonger Mn—O2 bond than the O1 site. Thus, the bondlength distortion is influenced not only by the Jahn—Tellereffect as Dent Glasser and Ingram (13) have pointed out butalso by the hydrogen bonding of the OH group.

For diaspore, the angle between the vectors O2—H andO2—O1 of 12° can be explained by repulsion of the H` byAl3` (14). A similar angle of 9(4)° was determined for grou-tite. If H would be on the line of oxygen centers, theMn—O2—H2 angles had to be 141°, in contrast to the ob-served 129(2)°. Groutite and ramsdellite (11, 13) are not only

t parameters are represented as 80% probability ellipsoids. Hydrogen is


isotopological but also crystallize in the same space group.The 2]1 structural channels in ramsdellite have an almostrectangular cross-section whereas corresponding H-bearingchannels in groutite are strongly distorted (Fig. 1). In addi-tion, the angular distortion of the OMn

3fragments is very

similar in both structures. From this point of view oneshould expect a topotactic transformation between the twostructures.

Structure of Manganite

Manganite, c-MnOOH, is a distorted derivative of pyro-lusite, b-MnO

2, isotypic with rutile (TiO

2). In the structure

of pyrolusite (space group P42/mnm), the MnO

6octahedron

is slightly elongated with two long 1.894(1) A_ and four short1.882(2) A_ Mn—O distances (25). Strongly elongated MnO

6octahedra were found in manganite (Fig. 9) where the direc-tion of elongation is the same as in pyrolusite. This en-hanced structural distortion is caused by an interaction ofthe Jahn—Teller effect (Mn4` is replaced by Mn3`) andhydrogen bonding. Two symmetrically distinct O sites (O1and O2) in manganite are three-coordinated by Mn3 .However, on the basis of the deviation of the OMn

3frag-

ments from planarity (Fig. 9), no specific O site can bepredetermined as preferred for OH. Both averages of OMn

3angles lead practically to the same value (115.5° for O1 and115.7° for O2).

To discuss the structural distortions and the orderedH arrangement in manganite, c-MnOOH, we derive this

FIG. 9. The Mn3` oxygen coordination in manganite. Atom displacemeattached to O1 at the apex of the octahedron.

structure from that of pyrolusite, b-MnO2

(space groupP4

2/mnm, a"4.4041, c"2.8765 A_ (25)). The observed cell

dimensions for manganite in the B121/d1 setting are a"

8.915(2), b"5.277(2), c"5.748(2) A_ , b"90.02(2)° (stan-dard space group setting is P12

1/c1). The above B-centered

cell is related to pyrolusite by a1:30-64*5%

"a/2.!/'!/*5%

,b1:30-64*5%

"b.!/'!/*5%

, and c1:30-64*5%

"c/2.!/'!/*5%

leading toan orthorhombic manganite subcell with a"4.458,b"5.277 and c"2.874 A_ . This subcell is sufficient to de-scribe the Jahn—Teller distortion due to octahedral Mn3` inmanganite and obeys the space group Pnnm (a nonisomor-phic subgroup of P4

2/mnm) with Mn at ..2/m (0, 0, 0) and

O at ..m (0.25, 0.375, 0). These coordinates lead to distortedMnO

6octahedra with two long 2.27 A_ and four short

1.93 A_ Mn—O bonds. The structural channels running par-allel to the c-axis may be described by edge-sharing chainsof empty distorted O

6octahedra which host the protons

necessary for charge balance. Corresponding subcells wereoriginally proposed for b-CrOOH and InOOH (26, 27) andfor high pressure polymorphs of NiOOH, FeOOH, VOOH,RhOOH, and ScOOH (28). If the subcell with space groupPnnm would describe the true symmetry of manganite andthe other polymorphs cited above, the H atoms must eitherreside at 0, 1

2, 0, leading to two symmetric O—H distances of

1.295 A_ , or H must be slightly displaced from the twofoldaxis (0#*x, 1

2#*y, 0) to form a covalent OH group with

an O—H distance of ca. 1 A_ and a strong hydrogen bond tothe opposite O of ca. 1.59 A_ . The latter arrangement wouldlead to H disorder. Brown (29) discusses the geometry of an

nt parameters are represented as 80% probability ellipsoids. Hydrogen is


OH···O bond and shows that with an O—O distance of ca.2.6 A_ a symmetric hydrogen bond is not expected to occur.To avoid H disorder, the structure must further degeneratein symmetry, thus the twofold axis along c is no longerallowed. In turn, either one of the nonisomorphic mono-clinic subgroups of Pnnm (P2

1/n11 or P12

1/n1, both P2

1/c,

standard setting) or the acentric orthorhombic subgroupsPn2

1m or P2

1nm are expected. b-CrOOH and InOOH both

crystallize in the latter orthorhombic subgroup (30, 10),while c-MnOOH prefers a monoclinic cell (Fig. 10). Formanganite, thus a cell setting different from the pseudo-pyrolusite cell has to be chosen in order to avoid a center ofsymmetry close to the supposed H position within thestructural channel. In this new cell, all atoms are on generalpositions and no split H positions occur within the struc-

FIG. 10. Octahedral structural model of manganite (a) and InOOH (b), sh1]1 channels.

tural channels. Nevertheless the high Pnnm pseudosym-metry still exists, giving rise to long-range proton disorderwhich was analyzed in terms of twinning. The twinneddomains are related to each other by a twofold axis parallelto the channel direction (c-axis in the pyrolusite cell or[101] in the manganite P2

1/c cell). The twinning can also be

described by a twin plane perpendicular to [101].

Behavior of Manganite and Groutite at ElevatedTemperatures

If manganite is heated in air, the a and b lengths of theB-centered cell increase to a stronger degree than thec length. This indicates that the hydrogen bonds which areoriented parallel to (001) are weakened in this plane. Under

owing unit cell outlines and hydrogen position (small spheres) in the endless


air flush at 300°C manganite oxidizes within minutes tob-MnO

2, pyrolusite, where the orientation of the octahedral

framework remains preserved. This observation confirmsprevious studies (31, 32). Due to the loss of H and oxidationto Mn4 , the b-axis decreases strongly (16%) from 5.277 to4.409 A_ . This change is accompanied by the loss of theJahn—Teller distortion and formation of square structuralchannels which were strongly distorted in manganite. Thecell dimensions of pyrolusite deviate significantly fromtetragonal symmetry which is caused either by strain due tothe transformation or by the formation of orthorhombicpyrolusite (21). Deviation from tetragonal symmetry is wellestablished for so-called secondary pyrolusite. On the basisof HRTM investigations it was proposed that manganiteoxidizes to an intergrowth of b-MnO

2and Mn

5O

8respon-

sible for the nontetragonal bulk properties (33). Champness(32) showed that the contraction of the b-axis during thetransformation from manganite to pyrolusite involvesformation of lamellar pores about 85 A_ apart parallel to(010) in B-cell setting. Yamada et al. (34) confirmed theseholes and also found ramsdellite domains in natural pyro-lusite which formed as an oxidation product of manganite.The above statement that in our experiment manganiteoxidized to pyrolusite should be taken with some caution.The recorded X-ray pattern is very diffuse, and the namepyrolusite was chosen because of very similar cell dimen-sions. Ripert et al. (35) have modeled the X-ray powderpattern of c-MnO

2by a random distribution of pyrolusite

and ramsdellite layers. Up to ca. 15% ramsdellite contribu-tion their calculated pattern is diffuse but still has thecharacteristics of pyrolusite.

The transformation of groutite to pyrolusite above 325°Cunder oxidizing conditions seems highly surprising becauseone would rather expect formation of the MnO

2modifi-

cation ramsdellite which is isotopological to groutite.However, previous studies (36—38) also obtained b-MnO

2(pyrolusite) as oxidation product. Klingsberg and Roy (39)reported that in air groutite persists indefinitely below130°C. At about 130°C it oxidizes partly to ramsdellitewithin two weeks; at 300°C it oxidizes within few hours.Above 300°C it transforms directly to pyrolusite. In a sim-ilar oxidation experiment of groutite at 300°C (36) addi-tional weak and diffuse single-crystal X-ray spots indicatedthe presence of metastable Mn

2O

3(hematite structure type)

intergrown with pyrolusite. The formation of pyrolusiteindicates that the linkage of octahedral chains is alteredleading to 1]1 channels which may be explained by twodifferent mechanisms. Lima-de-Faria and Lopes-Vieira (36)proposed a transformation mechanism where Mn migratesvia octahedral interstices, while the oxygen arrangementsremains preserved. In addition, slip planes responsible forthe transformation were suggested (40).

IR-spectroscopic investigation of ramsdellite, MnO2, has

disclosed a single crystallographically-ordered H2O mol-

ecule (21). Furthermore, ramsdellite used for a structuralinvestigation (11) contained 1.3 wt% H

2O. Thus it was

suggested that H2O is an integral part of the ramsdellite

structure (21). The high temperature transformation ofgroutite directly to pyrolusite without intermediate rams-dellite seems to confirm this assumption. However, asalready discussed for the manganite—pyrolusite transforma-tion, the existence of ramsdellite domains within pyrolusitecan not be excluded.

The transformation from groutite (»"139.5 A_ 3) to pyro-lusite (2»"111.6 A_ 3) is accompanied by a volume contrac-tion of 20%, whereas the transformation from manganite(»"135.2 A_ 3) to pyrolusite causes a volume contraction of17%. On the other hand, the transformation from groutiteto structurally related ramsdellite (»"120.4 A_ 3) causes avolume reduction of ca. 14%. There is strong evidence thatin transformations from MnOOH to MnO

2the close

packed oxygen framework does not remain preserved assuggested by Lima-de-Faria and Lopes-Vieira (36). Naturalpyrolusite formed due to transformation from manganiteexhibits (a) lamellar micropores 85 A_ apart parallel to (010)(32—34), (b) zipper-like domains of either ramsdellite (34) ortopotactic Mn

5O

8(33) causing specific pyrolusite X-ray

reflections to become diffuse. In contrast, manganite dis-plays very sharp X-ray reflections thus the existence ofsignificant groutite domains within the manganite precursorphase can be excluded. These observations are in betteragreement with a transformation model of submicroscopicslip planes where the structural strain, due to the volumedifference, is balanced by the formation of holes and rams-dellite (34) or Mn

5O

8(33) domains which both require

a smaller volume contraction. The evidence for a slip planemechanism even for the closely related structures of manga-nite and pyrolusite suggests a similar model for the morecomplex groutite—pyrolusite transformation.

Klingsberg and Roy (39) have discovered an intermediatephase, tentatively named ‘‘groutellite,’’ during the ramsdel-lite—groutite transformation. This phase with the possiblecomposition Mn

2O

3OH (orthorhombic, Pnma setting,

a"4.66, b"2.87, c"9.54 A_ ) was confirmed by later ex-periments (1, 41). However, it is assumed (1) that groutellitedoes not appear in samples reduced far from equilibriumand that it does not form during the reverse reaction, i.e.,oxidation of groutite to ramsdellite (39). The formation ofthis phase in the course of our groutite heating experimentsbelow 300°C can rather be excluded. Upon heating, groutiteX-ray reflections remained sharp and the intensity de-creased only slightly as expected from increased thermalvibration. The c-axis of groutite increased homogeneouslyfrom 10.665 A_ (25°C) to 10.773 A_ (325°C) whereas formationof intermediate ‘‘groutellite’’ should yield a significant de-crease of c to ca. 9.6 A_ . Formation of groutellite has recentlybeen suggested (1) during discharge of alkaline manganesedioxide batteries.


FTIR Spectroscopy on MnOOH Polymorphs

The subject of the following discussion are the peculiarabsorption features around 2000 cm~1 in the IR spectra ofmanganite and groutite. Even though they might be treatedas a sole, unique problem of MnOOH spectra, the authorsare convinced that only consideration of, and comparisonwith similar features in the spectra of comparable M3`

OOH (M3ÒOD) compounds [M"Al, Fe, Ga, Sc, dias-pore group (42); M"In, isotopological structure to manga-nite (30, 42); M"Co, Cr (43, 44)] provide a satisfactorysolution. Moreover, previous discussions of similar featuresin ‘‘A, B, C’’-type spectra (45, 46) of mostly organic substan-ces with medium to strong hydrogen bonds, provide addi-tional arguments. Thus, the band(s) around 2000 cm~1 inmanganite and groutite may be explained by one of thefollowing arguments which are grouped into three maincategories:

1. Both of the bands around 2600—2700 and&2000 cm~1 are O—H stretching bands. Their existencecan be explained in the following ways.

(A) Factor group splitting into two differently polar-ized components (see above) might lead to the two separ-ated absorption features. Counter argument: A splittingamount of &600 cm~1 of the two normal modes whichbelong to the same internal coordinate (i.e., in-phase andout-of-phase stretching) has never been observed yet.Rather, there is good evidence that the two species are veryclosely split, and both are contained in the broad 2600—2700 cm~1 band. This is also supported by the polarizationbehavior.

(B) Schwarzmann et al. (47) suggested proton tunnelingin a double minimum potential of the H-bond in manganite.Since the tunneling process, which may be described bydynamic H atom disorder, is in disagreement with thestructure refinements, they proposed tunneling stimulatedby the IR radiation during spectra acquisition. Counterargument: Even though a double minimum potential seemsvery likely in manganite (one minimum can be transformedinto the other one by a simple twin transformation) theirmodeling is in poor agreement with the wide separation ofthe two bands. Further, IR spectra collected at different (andalso very low) light intensities did not show any variations.Thus there is no evidence for a light-stimulated tunnelingprocess.

(C) Inelastic neutron scattering spectra where modeledassuming a highly symmetrical position of the H atom in thestructural channels of manganite (48). Similar suggestionswere published by Kamath and Ganguly (49), who de-scribed MnOOH as a ‘‘bronze’’ of H and MnO

2placing the

H atoms in the center of the structural channels. Counterargument: The peak positions of their studies suggest thatthey used different or altered material but pure manganite.Their findings are in strong disagreement with a previous (7)

and the present structure refinements, as well as with thefact that symmetrical H bonds are only observed atO—H···O distances below 2.5 A_ .

(D) There might be two slightly different environmentsaround the H atoms in manganite. The band positionsaround 2600—2700 and 2000 cm~1 indicate O—H···O dis-tances of &2.60 (actually observed) and &2.55 A_ (23). Thelatter might be produced by a distorted environment at thetwin boundaries in manganite. Counter argument: The al-most identical intensities of the two peaks are in disagree-ment with the comparably infrequent occurrence of the twinboundaries (refinement in a twin model instead of a disorderstructure!). A study with the IR-microscope using a 50]50 lm aperture yielded identical spectra in differently twin-ned sections of a 4]2]0.004 mm (100) manganite crystalsection.

In addition, the following observations might be con-sidered as a general counter argument of category 1: Upondeuteration (42) the two bands in manganite behave quitedifferently. Compared to the deuterated 2660 cm~1 band,the deuterated 2060 cm~1 band loses considerably intensity.The observed frequency shift of the 2660 cm~1 band(l

OH/l

OD"1.27) is in good agreement with the ‘‘stretching

frequency vs deuteration shift’’ correlation diagrams ofNovak (23). In contrast, the 2060 cm~1 band yields a shift of&1.34 which deviates considerably from the value of &1.1expected for short stretching frequencies. The observed shiftrather suggests a bending mode.

At low temperatures (82 K) the bands around 2600—2700 cm~1 shift to shorter wavenumbers (Table 5) which isthe usual behavior of stretching bands upon cooling. Thebands around 2000 cm~1 shift to higher wavenumberswhich is similar to the shift of the bending modes around1000—1150 cm~1. Finally, comparison with spectra of differ-ent M3ÒOH compounds (see above) suggests that the&2000 cm~1 bands are definitely not O—H stretchingmodes (detailed discussion below).

2. The two bands around 2600—2700 and &2000 cm~1

represent a single, broad OH stretching band interrupted bya minimum. This minimum can be explained in the follow-ing ways.

(A) The positions of the first overtones of the O—Hbending modes (i.e., two times the wavenumbers of the in-and out-of-plane bending modes at &1000 to 1150 cm~1)are in good agreement with the minimum between the2600—2700 and &2000 cm~1 bands. Thus, the transmission‘‘window’’ is caused by an overlap and Fermi resonance ofthese modes with the broad O—H stretching band (50).Counter argument: Due to the strong anharmonicity ofstrongly H-bonded O—H vibrations, the overtones must notbe calculated by a simple multiplication. The real overtonepositions are mostly shifted to lower wavenumbers (negativeanharmonicity constant X!) than expected by a such simplecalculation.


(B) Due to a very shallow potential surface, the groundstate of the O—H stretching vibration might be completelydepopulated by the IR radiation (comparable to laser ab-sorption studies with high photon flux), thus yielding a min-imum in the supposed broad band. Counter argument:Spectra acquired at different (and very low) light intensitiesdid not result in different gap widths or depths. If theground state is completely depopulated, absorptions atslightly lower wavenumbers (anharmonicity) due to thetransition from the first to the second state (‘‘hot bands’’)should be visible (51).

The appearance of more widely or closely separatedpeaks in comparable Me3ÒOH spectra or the occurrenceof very sharp component peaks at low temperatures may actas a general counter argument to the hypothesis statedabove. This is also confirmed by the different behavior of thetwo component bands at low temperatures or upon deuter-ation (see discussion above).

3. The 2600—2700 cm~1 bands are the true, fundamentalO—H stretching bands (related to the O—H···O distance of&2.60 A_ in the structures of manganite and groutite), andthe &2000 cm~1 band(s) are not stretching bands. How-ever, as the strong shift upon deuteration of manganiteshows, the latter are undoubtedly O—H modes.

(A) The bands around 2000 cm~1 represent additionalbending modes. Counter argument: Even in short H bondswith low-frequency stretching modes, the bending modes donot exceed 1400 cm~1. In addition, all bending modes al-lowed by factor group analysis have already been assignedto the 1000—1150 cm~1 group.

(B) The 2060 cm~1 band in manganite is a combina-tion band of the OH-stretching mode at 2660 cm~1 and theexcited lattice mode at 600 cm~1 (52). Thus, by 2660—600"2060, the position of the problematic band is accuratelyobtained. Counter argument: This combination works onlyin manganite, thus seems to be accidental. The contrarilyoriented shift of the two peaks at low temperatures is nolonger in agreement with this combination. According tonormal mode analysis (factor group analysis in a crystal) thecombination of an IR-active ‘‘u’’ mode (e.g., A

u, B

1u) with

another ‘‘u’’ mode always yields a ‘‘g’’ mode (IR-forbidden)in a centrosymmetric crystal (53).

(C) The &2000 cm~1 bands are combinations of theO—H stretching modes with low-wavenumber stretchingmodes (p) of the whole O—H···O bonds. Counter argument:Since the latter is in the wavenumber region around150 cm~1 it is rather suited to explain the broadening of thereal stretching bands (46) than the existence of the widelyseparated &2000 cm~1 bands.

(D) The bands around 2000 cm~1 might be ‘‘hotbands,’’ i.e. transitions from the excited state to the doublyexcited state of the stretching modes. Counter argument:A difference of &600 cm~1 between normal and ‘‘hot band’’can no longer be explained by anharmonicity. Hot bands

usually become more intense at higher temperatures. This isopposite to the present observations.

(E) The &2000 cm~1 bands represent combinationsof an IR-active and Raman-active O—H bending mode.Counter argument: This is a weak argument, since theRaman modes cannot be proven due to the strong lightabsorption of the strongly colored MnOOH minerals. How-ever, comparable Raman spectra of diaspore (42) did notsupport this hypothesis.

(F) The bands around 2000 cm~1 are the first overtoneof the bending modes at 1000—1150 cm~1. Counter argu-ment: According to normal mode analysis (see above) thefirst overtones are IR-forbidden (2]‘‘u’’"‘‘g’’). Further, theintensities (compared to the fundamental stretching andbending modes) are much too strong for an overtone.

Even though at least one counter argument can be foundfor each single hypothesis, the most probable explanation isgiven by group 3, which lacks a general counter criticism.Among this group, explanation 3F, an overtone of bendingmodes, is the most reasonable one. However, how can thecounter arguments be handled seriously?

A thorough consideration of the behavior of medium tostrong hydrogen bonds leads inevitably to the field of an-harmonicity. A general review of this topic is given by Hadziand Bratos (46) and Sandorfy (54). Which facts support theintimate relation of the &2000 cm~1 bands to anharmonic-ity phenomena?

(a) The bands are strong in manganite and groutite butbecome less intense in diaspore, goethite, etc., which showweaker H bonding. It is well known that the anharmonicityof H bonds is generally very strong. Moreover, it increasesdramatically with decreasing O—H···O distance and O—Hstretching frequency (55).

(b) Compared to the real stretching bands around2600—2700 cm~1, the &2000 cm~1 bands become muchmore intense at low temperatures. It is well proven that theanharmonicity increases with decreasing O—H···O distanceat low temperature (54).

(c) In contrast, the &2000 cm~1 band becomes muchweaker and narrow upon deuteration. In M3ÒOD com-pounds with weaker H bonds, the band disappears almostcompletely. It is well understood that the anharmonicitydecreases with replacement of a constituent light atom bya heavier one (H/D).

(d) The position around 2000 cm~1, which is lower thanthe expected values (2]bending fundamental), confirms thestrong anharmonicity (strong, negative anharmonicity con-stant X

OH(55)).

The counter arguments turn into supporting arguments ifwe consider the above mentioned facts. Moreover, in thecase of strong anharmonicity, the overtone (forbidden by nor-mal mode analysis) should appear according to local modeanalysis. Normal mode analysis is only valid for harmonicvibrations or for cases in which a weak anharmonicity can


be treated by a simple perturbation theory. In the case ofstrong anharmonicity the normal mode analysis has to bereplaced by local mode analysis (56, 57) which proved versa-tile in the interpretation of overtone spectra.

Finally, the high intensity of the overtone bands may beexplained by anharmonicity resonance interaction (this seemsto be a better term than ‘‘Fermi’’ resonance) with the main,broad O—H stretching band (46). The intensity of resonanceincreases with increasing H bond strength (increasing an-harmonicity) and with approaching band positions. In addi-tion, there is good evidence that the high-energy shoulder orpeak (at low temperatures in groutite) around 2900 cm~1 isthe resonance-enhanced second overtone of the bendingmodes under discussion. The peculiar, asymmetrical shapesof the overtone bands provide an additional supportingargument for resonance-enhanced intensities.

ACKNOWLEDGMENTS

This study was supported by the Swiss ‘‘Nationalfonds’’. Thanks are dueto R. Oberhansli for electron microprobe analysis of groutite, Ch. Hoff-mann for help during the course of structure refinement, and M. Kunz forvaluable discussions. We are indebted to G. R. Rossman for supporting thisstudy with his FTIR equipment, for providing the diaspore samples, andfor numerous helpful discussions. Financial support for the FTIR equip-ment came from the National Science Foundation, Grant EAR 9218980.E.L. is indebted to the ‘‘Fonds zur Forderung der wissenschaftlichenForschung, Austria’’ for financial support during an ‘‘Erwin-Schrodinger’’fellowship, Project J01098-GEO.

REFERENCES

1. Y. Chabre and J. Pannetier, Prog. Solid State Chem. 23, 1—130 (1995).2. P. M. de Wolff, Acta Cryst. 12, 341—345 (1959).3. L. A. H. MacLean and F. L. Tye, J. Solid State Chem. 123, 150—160

(1996).4. M. H. Rossouw, D. C. Liles, and M. M. Thackeray, Mat. Res. Bull. 27,

221—230 (1992).5. M. J. Garrido, Bull. Soc. Franc7 . Mine& ral. 58, 224—241 (1935).6. M. J. Buerger, Z. Krist. 95, 163—174 (1936).7. H. Dachs, Z. Krist. 118, 303—326 (1963).8. J. V. Smith, ‘‘Geometrical and structural crystallography’’ (J. V. Smith,

Ed.), pp. 326—329. John Wiley & Sons, New York, 1982.9. Y. D. Kondrashev and A. I. Zaslavsky, Izv. Akad. Nauk SSSR, Ser. Fiz.

15, 179—186 (1951).10. M. Pernet, C. Berthet-Colominas, M. A. Alario-Franco, and A. N.

Christensen, Phys. Status Solidi A 43, 81—88 (1977).11. A. M. Bystrom, Acta Chem. Scand. 3, 163—173 (1949).12. J. W. Gruner, Am. Mineral. 32, 654—659 (1947).13. L. S. Dent Glasser and L. Ingram, Acta Crystallogr., Sect. B 24,

1233—1236 (1968).14. W. R. Busing and H. A. Levy, Acta Crystallogr. 11, 798—803 (1958).15. ‘‘Structure determination package (SDP).’’ Enraf Nonius, Delft, The

Netherlands, 1983.16. G. M. Sheldrick, ‘‘SHELX93, Program for crystal structure determina-

tion.’’ University of Gottingen, Germany, 1993.17. Siemens, ‘‘XLS/SHELXTL PC, Release 4.1.’’ Siemens Analytical X-

Ray Instruments, Inc., Madison, WI, 1991.18. R. D. Shannon, M. A. Subramanian, A. N. Mariano, T. E. Gier, and

G. R. Rossman, Am. Mineral. 77, 101—106 (1992).

19. E. Libowitzky and G. R. Rossman, Am. Mineral. 81, 1080—1091 (1996).20. E. Schwarzmann and H. Marsmann, Z. Naturforsch. 21B, 924—928

(1966).21. R. M. Potter and G. R. Rossman, Am. Mineral. 64, 1199—1218

(1979).22. D. M. Adams and D. C. Newton, ‘‘Tables for factor group and point

group analysis.’’ Beckman-RIIC Limited, Croydon, England, 1970.23. A. Novak, Struct. Bonding 18, 177—216 (1974).24. A. Szytula, A. Burewicz, Z. Dimitrijewic, S. Krasnicki, H. Rzany, J.

Todorovic, A. Wanic, and W. Wolski, Phys. Status Solidi 26, 429—434(1968).

25. A. A. Bolzan, C. Fong, B. J. Kennedy, and C. J. Howard, Aust. J. Chem.46, 939—944 (1993).

26. A. N. Christensen, Acta Chem. Scand. 18, 1261—1266 (1964).27. A. N. Christensen, Inorg. Chem. 5, 1452—1453 (1966).28. J. Chenavas, J. C. Joubert, J. J. Capponi, and J. Marezio, J. Solid State

Chem. 5, 1—15 (1973).29. I. D. Brown, Acta Crystallogr. Sect. A 32, 24—31 (1976).30. M. S. Lehmann, F. K. Larsen, F. R. Poulsen, A. N. Christensen, and

S. E. Rasmussen, Acta Chem. Scand. 24, 1662—1670 (1970).31. D. R. Dasgupta, Min. Mag. 35, 131—139 (1965).32. P. E. Champness, Min. Mag. 38, 245—248 (1971).33. J. H. Rask and P. R. Busek, Am. Mineral. 71, 805—814 (1986).34. N. Yamada, M. Ohmasa, and S. Horiuchi, Acta Cryststallogr. Sect.

B 42, 58—61 (1986).35. M. Ripert, C. Poinsignon, Y. Chabre, and J. Pannetier, Phase

¹ransitions 32, 205—209 (1991).36. J. Lima-de-Faria and A. Lopes-Vieira, Min. Mag. 33, 1024—1031

(1964).37. L. S. Dent Glasser and I. B. Smith, Min. Mag. 35, 327—334 (1965).38. R. Giovanoli, and U. Leuenberger, Helv. Chim. Acta 52, 2333—2347

(1969).39. C. Klingsberg and R. Roy, Am. Mineral. 44, 819—838 (1959).40. B. G. Hyde and S. Andersson, ‘‘Inorganic crystal structures’’

(B. G. Hyde and S. Andersson, Eds.). John Wiley & Sons, New York,1989.

41. L. S. Dent Glasser and I. B. Smith, Min. Mag. 36, 976—987 (1968).42. E. Schwarzmann, and H. Sparr, Z. Naturforsch. 24B, 8—11 (1969).43. R. G. Delaplane, J. A. Ibers, J. R. Ferraro, and J. J. Rush, J. Chem. Phys.

50, 1920—1928 (1969).44. R. G. Snyder and J. A. Ibers, J. Chem. Phys. 36, 1356—1360 (1962).45. Y. I. Ryskin, ‘‘The infrared spectra of minerals’’ (V. C. Farmer, Ed.),

pp. 137—181. Mineralogical Society, London, 1974.46. D. Hadzi and S. Bratos, ‘‘Structure and spectroscopy’’ (P. Schuster, G.

Zundel, and C. Sandorfy, Eds.), Vol. II, pp. 565—611. North-Holland,Amsterdam, 1976.

47. E. Schwarzmann, O. Glemser, and H. Marsmann, Z. Naturforschung21B, 1128—1131 (1966).

48. F. Fillaux, C. H. Cachet, H. Ouboumour, J. Tomkinson, C. Levy-Clement, and L. T. Yu, J. Electrochem. Soc. 140, 585—591 (1993).

49. P. V. Kamath and S. Ganguly, Mater. ¸ett. 10, 537—539 (1991).50. M. F. Claydon and N. Sheppard, Chem. Commun. 23D, 1431—1433

(1969).51. H. Graener, J. Phys. Chem. 95, 3450—3453 (1991).52. C. Cabannes-Ott, Ann. Chim. Ser. 13, 5, 905—960 (1960).53. E. B. Wilson, J. C. Decius, and P. C. Cross, ‘‘Molecular vibrations’’

(E. B. Wilson, J. C. Decius and P. C. Cross, Eds.), p. 388. DoverPublications, Inc., New York, 1980.

54. C. Sandorfy, ‘‘Structure and spectroscopy’’ (P. Schuster, G. Zundel, andC. Sandorfy, Eds.), Vol. II, pp. 613—654. North-Holland, Amsterdam,1976.

55. H. D. Lutz and M. Schmidt, Spectrosc. Acta 47A, 585—594 (1991).56. M. S. Child and L. Halonen, Adv. Chem. Phys. 57, 1—58 (1984).57. I. M. Mills and A. G. Robiette, Mol. Phys. 56, 743—765 (1985).

Documents

Hydrogen Bonding and Jahn–Teller Distortion in Groutite,α-MnOOH, and Manganite,γ-MnOOH, and Their Relations to the Manganese Dioxides Ramsdellite and Pyrolusite